DESCRIPTION: Dr. Marmorstein proposes to study the mode of DNA recognition by five fungal transcriptional activator proteins that bind to DNA as dimers using binuclear Zn2Cys6 clusters. DNA complexes of two were solved by him as a postdoctoral in Steve Harrison's laboratory at Harvard: GAL4 and PPR1. These bind by interacting each binuclear Zn cluster with a CGG trimer, the two trimers being separated by 11 other base pairs for GAL4 or 6 base pairs for PPR1. Each protein contains an N-terminal DNA-binding domain with the Zn cluster, a hydrophobic region of repeated heptads that is believed to be responsible for dimerization of the protein, and a longer transcriptional activator domain that can involve several hundred amino acids. All of the crystallographic studies, past and planned, involve only the isolated N-terminal DNA-binding and dimerization heptad domains. Clusters for the above proteins are arranged antisymmetrically; that is, the consensus binding sequence is CGGnxCCG. Other fungal transcription activators use a reversed sequence from the above, GCCn6GGC, or one in which the two CGG sites are repeated in tandem rather than antiparallel: CGGn5CGG. At issue in this work is the question of how each protein dimer positions its two Zn clusters in space so they can sense the length of DNA duplex separating the two C/G contact sites. These proteins read DNA sequence information in two stages: the first is the detection of acceptable C/G binding sites for the Zn clusters, and the second is the spacing between the two clusters. Dr. Marmorstein will continue the 2.7 Angstrom resolution GAL4 study begun at Harvard by solving the structure of a new crystal form that diffracts to better than 2.0 Angstrom. No further work is planned on PPR1. Two new factors that bind in the anitparallel mode like GAL4 will be solved, PUT3 and qa-1F. Preliminary crystals are in hand that diffract to 3.2 Angstrom (PUT3) and 6 Angstrom (qa-lF), with further crystallization screening under way. Dr. Marmorstein also plans to examine two factors that bind in a different manner. LEU3 binds to an inverted site in which G and C are exchanged, GCCn6 GGC and HAP1 binds to what appears to be the GAL4 sequence with CGG in tandem rather than opposed: CGGnnTAnCGG. Crystallization trials are under way for both proteins, with small crystals of HAP1 already. Information from this project will afford Dr. Marmorstein a unique opportunity to address an important problem in DNA recognition: the role of spatial geometry as a recognition component, and how this is engineered into the structure of the recognition protein. The dimerization heptads that hold the Zn "reading heads" in the proper geometry will be of more interest than simple Zn cluster binding itself, and the cumulative set of structures for comparison will be far more interesting than individual structures would have been. The whole of this project will be distinctly greater than the sum of its parts.